See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/41192564

Modeling of Tool-Tissue Interactions for

Computer-Based Surgical Simulation: A Literature
Review

Article in Presence Teleoperators & Virtual Environments October 2008

DOI: 10.1162/pres.17.5.463 Source: PubMed

CITATIONS READS

97 126

3 authors:

Sarthak Misra K.T. Ramesh

University of Twente Johns Hopkins University
144 PUBLICATIONS 1,448 CITATIONS 262 PUBLICATIONS 6,775 CITATIONS

SEE PROFILE SEE PROFILE

Allison Okamura
Stanford University
295 PUBLICATIONS 6,978 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

X-ray phase-contrast imaging studies of dynamic fracture in ceramic materials View project

All content following this page was uploaded by Allison Okamura on 03 December 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
Sarthak Misra Modeling of Tool-Tissue
K. T. Ramesh
Allison M. Okamura* Interactions for Computer-Based
125 Computational Science and
Engineering Building
Surgical Simulation: A Literature
Department of Mechanical Review
Engineering
The Johns Hopkins University
3400 North Charles Street
Baltimore, MD 21218
Abstract

Surgical simulators present a safe and potentially effective method for surgical train-
ing, and can also be used in robot-assisted surgery for pre- and intra-operative plan-
ning. Accurate modeling of the interaction between surgical instruments and organs
has been recognized as a key requirement in the development of high-fidelity surgi-
cal simulators. Researchers have attempted to model tool-tissue interactions in a
wide variety of ways, which can be broadly classified as (1) linear elasticity-based,
(2) nonlinear (hyperelastic) elasticity-based finite element (FE) methods, and (3)
other techniques not based on FE methods or continuum mechanics. Realistic mod-
eling of organ deformation requires populating the model with real tissue data
(which are difficult to acquire in vivo) and simulating organ response in real time
(which is computationally expensive). Further, it is challenging to account for con-
nective tissue supporting the organ, friction, and topological changes resulting from
tool-tissue interactions during invasive surgical procedures. Overcoming such obsta-
cles will not only help us to model tool-tissue interactions in real time, but also en-
able realistic force feedback to the user during surgical simulation. This review paper
classifies the existing research on tool-tissue interactions for surgical simulators spe-
cifically based on the modeling techniques employed and the kind of surgical opera-
tion being simulated, in order to inform and motivate future research on improved
tool-tissue interaction models.

1 Introduction

People have always sought ways to understand and model the structure
and function of the human body. The earliest known anatomical models used
for surgical planning were recorded around 800 BC in India for the procedure
of rhinoplasty, in which a flap of skin from the forehead is used to reconstruct
a nose. These ancient practitioners used leaves to represent three-dimensional
flexible tissues and plan the surgical operation (Carpue, 1981). Before the ad-
vent of medical imaging a century ago, the only practical way to see inside the
human body was to observe an operation or by dissection. However, cultural
and religious beliefs, the difficulty of obtaining cadavers, and lack of refrigera-

Presence, Vol. 17, No. 5, October 2008, 463 491

2008 by the Massachusetts Institute of Technology *Correspondence to [email protected].

Misra et al. 463

464 PRESENCE: VOLUME 17, NUMBER 5

tion imposed restrictions on the widespread use of dis-

section. Frustrated by these limitations, Louis Thomas
Jerome Auzoux, a nineteenth century French physician,
improved and popularized anatomical papier mache
models (Davis, 1975). By the early twentieth century,
inexpensive and realistic plastic anatomical models that
could be assembled and painted became popular with
medical students. Human cadavers and animals, how-
ever, are still used for training and, in some cases, surgi-
cal planning.
In the past decade, advances in computer hardware
and software, and the use of high-fidelity graphics, have
made it possible to create simulations of medical proce-
dures. Computer-based surgical simulation provides an
efficient, safe, and ethical method for training clinicians Figure 1. Hysteroscopy training simulation environment coupled
by emphasizing the users real-time interaction with with a haptic device (Harders et al., 2006). Reprinted from Studies in
Health Technology and Informatics, Vol. 119, Harders et al. Highly
medical instruments, surgical techniques, and realistic
realistic, immersive training environment for hysteroscopy, pp. 176
organ models that are anatomically and physiologically
181, (2006), with permission from IOS Press.
accurate. The objective is to create models that support
medical practitioners by allowing them to visualize, feel,
and be fully immersed in a realistic environment. This
implies that the simulator must not only accurately rep-
eters, realistic modeling and simulation of tissue me-
resent the anatomical details and deformation of the
chanics is an active research area. There is a large vol-
organ, but also feed back realistic tool-tissue interaction
ume of literature on the topic of tissue modeling that is
forces to the user. As an example, consider the proto-
distributed across the biomechanics, robotics, and com-
type hysteroscopy training simulator shown in Figure 1,
puter graphics fields. For certain topics outside the
which was developed at Eidgenossische Technische
scope of this review, good references are available in the
Hochschule (ETH) Zurich (Harders et al., 2006).1 This
literature:
surgical simulator allows real-time visualization of the
surgical procedure along with force feedback to the There exists a rich literature in the biomechanics
user. community involving the measurement of tissue
The development of realistic surgical simulation sys- properties of specific organs. In the 1970s, re-
tems requires accurate modeling of organs/tissues and searchers such as Fung (1973, 1993) and Yamada
their interactions with the surgical tools. The benefits of (1970) applied the techniques of continuum me-
tissue modeling are not only useful for training, plan- chanics to soft tissues and conducted extensive tests
ning, and practice of surgical procedures, but also opti- to characterize tissue properties.
mizing surgical tool design, creating smart instru- Extensive work has been done by researchers in the
ments capable of assessing pathology or force-limiting area of computer graphics to model and simulate
novice surgeons, and understanding tissue injury mech- deformable bodies in real time (Gibson & Mirtich,
anisms and damage thresholds. Given the complexity of 1997). The present survey does not cover work in
human organs and challenges of acquiring tissue param- which the focus is to produce seemingly realistic
visualization effects of deformation, while ignoring
1
Color versions of the figures in this article accompany the online
the physics underlying tissue deformation.
article, available at http://www.mitpressjournals.org/toc/pres/17/5. Approaches for complete simulator design, specific
 Misra et al. 465

Table 1. Survey of tool-tissue interaction models for surgical simulation and robot-assisted surgery*

Operation Model

Linear elastic Visco-hyperelastic

FE Hyperelastic FE FE Other methods

Deformation Simulation only Basdogan et al. Y. Liu et al. Puso and Weiss LEM (2002; 2002)
(via indentation) Bro-Nielsen X. Wu et al. Mass- (1994; 1996; 1996)
Cotin et al. spring- (2003; 1992)
Frank et al. damper model (2000)
Gladilin et al. Thin-walled model (1999)
Kuhnapfel et al. PCMFS (2005) and
PAFF (2006)
BEM solution (2001)
J-shaped function (1994)

Real/phantom Gosline et al. Carter et al. Szekely et al. Curve-fitting (2003)

tissue studies Kerdok et al. Chui et al. Miller Greens function (2002)
Sedef et al. Davies et al. Nava et al.
viscoelastic Hu and Desai Kim et al.
Molinari et al.

Rupture Simulation only Alterovitz et al. Nienhuys and Spring model (2004)
(via needle van der Volume model (2001)
insertion) Stappen 3D chain mail (2004)

Real/phantom Crouch et al. Resistive-(1997)

tissue studies DiMaio and force model (2000)
Salcudean Curve-fitting (2001)
Hing et al. Friction model (2004)

Cutting Simulation only Picinbono et al. Picinbono et al. Hybrid model (2000)
(via blade/ W. Wu and Heng
scissors)

Real/phantom Chanthasopeephan Fracture mechanics (2001)

tissue studies et al. Haptic Scissors (2003)

*Within each category, authors are listed alphabetically and simulators that provide haptic feedback are designated by .

medical applications, and training evaluation meth- provides a mathematical framework to model the defor-
ods have also been widely studied in the last two mation of biological tissues, the primary focus of this
decades (A. Liu, Tendick, Cleary, & Kaufmann, review is to compile the research done in the area of
2003; Satava, 2001). modeling and simulating surgical tool-tissue interactions
This paper summarizes the literature and presents in real time using the principles of continuum mechan-
a taxonomy (Table 1) of significant work in the field of ics and finite element (FE) methods, respectively. Some
realistic modeling of tool-tissue interactions for simulation other methods that could be used to simulate physically
and robot-assisted surgery. Since continuum mechanics realistic tool-tissue interactions are also discussed.
466 PRESENCE: VOLUME 17, NUMBER 5

This paper is organized as follows. Section 2 summa- where is the gradient operator with respect to x. If
rizes the basic concepts and theories of linear and non- the displacement is u, then y x u. Thus,
linear elasticity, while Section 3 provides an overview of
F I u, (2)
the FE modeling technique. Sections 4 and 5 classify
the prior research work that has been done in modeling where I is the identity tensor. A useful measure of strain
noninvasive and invasive surgical tool-tissue interac- is the Green strain tensor, E, defined by
tions, respectively. Realistic tool-tissue interaction mod-
els require populating models with accurate material 1
E FTF I. (3)
properties, and Section 6 provides a summary of some 2
of the methods for acquiring tissue properties. Section 7
Substituting Equation 2 in Equation 3 results in
lists some of the commercially available surgical simula-
tors. Finally, Section 8 concludes by providing some 1
E u uT uT u. (4)
important directions for research in the area of realistic 2
modeling of tool-tissue interactions.
In linear elasticity theory, strains are assumed to be
small (u 1), hence,
2 Continuum Mechanics for Tissue 1
Modeling E u uT, (5)
2

The study of deformation or motion of a continu- where is the infinitesimal strain tensor. Thus, linear
ous material under the action of forces is known as con- elasticity theory is valid only for small strains (12%).
tinuum mechanics. The objective of this section is to One of the fundamental drawbacks of using a linear
provide a brief introduction to the mechanics of soft elasticity formulation to describe soft tissues is that sur-
tissues using the theories of linear and nonlinear elas- gically relevant strains often significantly exceed the
ticity. small strain limit, invalidating the assumption of linear-
The field equations of continuum mechanics are nor- ity. However, it is used in many simulation applications
mally formulated using tensors. An overview of tensor due to analytical simplicity and computational efficiency,
analysis is beyond the scope of this paper, but Ogden so it is described next.
(1984) provides a good introduction to this subject and
its application to continuum mechanics. Tensor nota- 2.2 Linear Elasticity
tions and manipulations consistent with the mechanics
literature are used in the derivations in this paper; for Linear elastic modeling of soft tissues is the most
example, boldface characters signify tensors, matrices, widely used approach within the robotics and haptics
and vectors, while normal characters are scalar quanti- community. Materials exhibiting linear elasticity obey
ties. the generalized Hookes Law, which relates the stresses,
, and infinitesimal strains, , by the tensor of elastic
moduli, C, as
2.1 Kinematics of Continua
C:. (6)
Consider a body undergoing deformation so that
material points initially at x are mapped to spatial loca- where : denotes double contraction, is also known as
tions, y. Then the deformation gradient tensor, F, is the Cauchy stress tensor and Equation 6 could be re-
defined by written as ij m1
3
n1
3
Cijmnmn, where Cijmn is a
fourth-order tensor with 81 constants that are specific
F y, (1) to the material. The subscript indices represent the
 Misra et al. 467

Figure 2. Tissue fiber orientation of the heart on the inside surface, (a) and (b), and outside surface, (c) and (d), constructed using diffusion
tensor imaging (Zhukov & Barr, 2003). Images are printed with permission from IEEE 2003.

components of the stress and strain tensor. Symmetry of 2.2.1 Linear Viscoelasticity. Most soft tissues
the stress and strain tensors leads to Cijmn Cjimn and are inherently viscoelasticthey have a response based
Cijmn Cjinm, and the postulated existence of a strain on both position and velocity. Viscoelastic materials ex-
energy density leads to Cijmn Cmnij. Thus, Cijmn has hibit properties of both elastic solids and viscous fluids.
21 independent constants (called moduli) for a fully Similar to elastic materials, linear viscoelastic materials
anisotropic material, that is, a material whose properties retain the linear relationship between stress and strain,
change with direction. but the effective moduli depend on time. For small
If the assumption is made that the material is isotro- strains, the general linear viscoelastic constitutive equa-
pic, then the material properties can be described by just tions can be derived by separating the stresses and
two independent parameters: Youngs modulus, E, and strains into the hydrostatic (superscript H) and devia-
Poissons ratio, . E and are related to the shear mod- toric (superscript D) components:
ulus, ,
H D (8)
E
. (7) H D (9)
21
Hydrostatic stresses/strains act to change the volume of
In the vast majority of surveyed literature (the research
the material, but maintain shape, while deviatoric or
cited under Linear elastic FE in Table 1), E and are
shear stresses/strains are those that distort the shape,
the two parameters used to describe the soft tissue
but preserve volume (in isotropic and linear elastic ma-
properties. Most biological materials are, however, in-
terials). The hydrostatic stresses and strains are related
trinsically anisotropic. For example, a soft tissue con-
by
taining fibers aligned along an axis will have different
properties along and transverse to that axis. A summary H 3KH, (10)
of such anisotropies is presented in Spencer (1972) and
Figure 2 depicts the orientation of muscle fibers in the and the bulk modulus, K E/3(1 2), has small
heart (Zhukov & Barr, 2003), while the nonlinear elas- variations with time as compared to the shear modulus
tic behavior of myocardial tissue and its application to and hence, K is considered to be independent of time.
surgical simulation is highlighted in Misra, Ramesh, and However, the deviatoric stresses and strains can be re-
Okamura (2008). lated by
468 PRESENCE: VOLUME 17, NUMBER 5

Figure 3. Examples of characteristic properties of viscoelastic materials: (a) creep and creep recoveryfor a constant applied shear stress 0
results in an increase in shear strain; (b) stress relaxationfor a constant applied shear strain 0 results in a decrease in shear stress until it
reaches a steady-state value.

where J (t) is the creep compliance for constant stress,

iD

jD
N M

pi i qj , (11) 0, and G (t) is the stress relaxation modulus for con-

t t j
i0 j0 stant strain, 0. The creep compliance and relaxation
modulus are empirically determined and describe the
where pi and qj are material constants. In Equation 11, creep and stress relaxation behavior of the viscoelastic
the indices N and M depend on the number of material material as a function of time.
constants required to have good fit with the experimen- In many cases, viscoelastic properties of soft tissues
tal results. are represented by rheological models, which are ob-
Two characteristic behaviors specific to viscoelastic tained by connecting springs (elastic elements) and
materials are creep and stress relaxation. Creep occurs dashpots (viscous elements) in serial or parallel combi-
when a constant stress applied to the material results in nations (Fung, 1993). Three simple material models
increasing strain. On the other hand, stress relaxation used to represent solids are the Maxwell, Kelvin-Voigt
occurs when a material is under constant strain but the (or Voigt), and Zener standard linear solid (or Kelvin)
stress decreases until it reaches some steady-state value. models shown in Figures 4(a), (b), and (c), respectively.
Figures 3(a) and 3(b) depict the behaviors of creep and For the Maxwell model, the relaxation behavior is
stress relaxation, respectively. The one-dimensional acceptable, but creep behavior is insufficiently modeled.
models of creep deformation and stress relaxation are The creep compliance and stress relaxation modulus for
the Maxwell model are given by
(t) 0Jt (12)
1 t
Jt , (14)
t 0Gt, (13) k b
 Misra et al. 469

(Fung, 1993). A hyperelastic material is characterized

by the existence of a strain energy density function,
W(F). The stress in the material as a result of deforma-
tion can be obtained from

WF
P , (19)
F

where P is the first Piola-Kirchhoff stress tensor and F is

the previously defined deformation gradient tensor. The
Cauchy stress tensor and first Piola-Kirchhoff stress ten-
sor are related by
Figure 4. Standard viscoelastic models commonly used to represent
soft tissues (a) Maxwell (b) Kelvin-Voigt (or Voigt) (c) Zener standard
PF T J, (20)
linear solid (or Kelvin) (Fung, 1993).
where J det (F). There are several formulations for
the strain energy density function, for example, the St.
tk
Venant-Kirchhoff, Blatz-Ko, Ogden, Mooney-Rivlin,
Gt ke b . (15) and Neo-Hookean models (Ogden, 1984). Ogden and
Mooney-Rivlin strain energy density formulations pro-
where t is time. The Voigt model provides a satisfactory
vide a fairly accurate representation of the constitutive
first-order approximation of creep, but is an inadequate
laws for many biological tissues (Holzapfel, 2000). In
model for stress relaxation. For the Voigt model, the
an Ogden model, the strain energy density function for
governing equation is that of an elastic material, so
an isotropic material is given in terms of the principal
there is no relaxation of stress, hence, the creep compli-
stretches, i, as
ance is given by

N
1 tk k k
Jt 1 e b . (16) W 1 2k 3k 3, (21)
k k1
k

The Zener standard linear solid model provides a good

where 123 1, i.e. thermal incompressibility, and k
qualitative description of both creep and stress relax-
and k are material parameters determined from experi-
ation. The creep compliance and relaxation modulus are
ments. The Mooney-Rivlin model commonly used to
given by
represent rubber-like materials, is widely used for soft

Jt
1

1
k1 k2
tk2
1e b , (17)
tissues and is given in terms of the principal invariants,
Ii, for isotropic and incompressible materials as

Gt
k1
k1 k2 tk1k2
k2 k1e b . (18)
W C1I1 3 C2I2 3,

where C1 and C2 are material constants. The principal

(22)

invariants are defined in terms of right Cauchy-Green

2.3 Nonlinear Elasticity tensor, C FTF, as

Elastic materials undergoing deformations with I 1 C:I, (23)

large strains (12%) are described by nonlinear elastic-
ity theory. In order to model biological tissues, it is 1
I2 C:I2 C:C, (24)
common to use hyperelasticity and visco-hyperelasticity 2
470 PRESENCE: VOLUME 17, NUMBER 5

I 3 det C. (25) Two adjacent elements are connected via nodes. The
elastic behavior of each element is categorized using
Some researchers have used the Neo-Hookean model to matrices in terms of the elements material and geomet-
represent soft tissues. The Neo-Hookean strain energy ric properties, and distribution of loading within the
density function is a special case of the Mooney-Rivlin element and at the nodes of the element. Linear or qua-
model and is given by dratic shape or interpolation functions are used to ap-
proximate the behavior of the field variables at the
W C1I1 3. (26)
node. The element behavior is characterized by partial
If the material parameter constants for Ogden or differential equations governing the motion of material
Mooney-Rivlin models are defined in terms of creep and points of a continuum, resulting in the following dis-
stress relaxation functions, then the material can be crete system of differential equations:
modeled as visco-hyperelastic, which may represent the
realistic behavior of many soft tissues (Fung, 1993). Mu Cu Ku F R, (27)

where M, C, and K are the element mass, damping, and

3 Finite Element Modeling stiffness matrices, respectively, u is the vector of nodal
displacements, and F and R are the external and internal
The FE method is a numerical technique for solv- node force vectors, respectively. All these matrices and
ing field equations, typically partial differential equa- vectors may be time dependent. One approach to solve
tions, and has been used in the last decade to simulate Equation 27 in a quasi-static manner is by setting u
soft tissue deformation by solving the equations of con- u 0. Thus, with every simulation iteration, a large
tinuum mechanics. The FE method originated from the number of element-stiffness and element-force vectors
need to find approximate solutions to complex prob- are assembled, which leads to a system of algebraic
lems in elasticity and vibration analysis (Cook, Malkus, equations, called the global system. The accuracy and
& Plesha, 1989). Its development can be traced back to numerical efficiency of the FE method lies largely in the
the work by Hrennikoff (1941) and Courant (1942) development of effective pre- and post-processors, and
(Fung & Tong, 2001). Over the years, the FE method algorithms for efficiently solving large systems of equa-
has spread to applications in many different areas of en- tions. On the other hand, Equation 27 can be solved
gineering, including structural analysis in civil and aero- with a dynamic approach, using implicit or explicit inte-
nautical engineering, thermal analysis, and biomechan- gration schemes. For explicit integration methods, the
ics. Numerous FE computer programs are commercially state at a given instant is a function of the previous time
available for general or specific applications. These in- instants, while for the implicit scheme, the state at a cer-
clude ABAQUS (Simulia), ADINA (ADINA R & D tain instant cannot be explicitly expressed as a function
Inc.), ANSYS (ANSYS Inc.), DYNA3D (Lawrence Liv- of the state at the previous time step. The implicit
ermore National Laboratory), FEMLAB (COMSOL scheme involves inversion of the stiffness matrix at each
Inc.), GT STRUDL (Georgia Tech-CASE Center), NX time step, typically a computationally expensive process.
I-deas (Siemens PLM Software), and NASTRAN (MSC The explicit scheme can be easily implemented by avoiding
Software Corp.). This section provides a very high-level the matrix inversion process but suffers from numerical
overview of the FE method; there are numerous textbooks instability under an inappropriately chosen integration
that deal with this subject, e.g., Cook et al. (1989), Desai time step. Thus, explicit time integration is only condition-
and Abel (1972), and Zienkiewicz, Taylor, and Zhu (2005). ally stable, potentially requiring very small time steps to
In the FE method, the continuum is divided or provide a suitably accurate and stable solution. Properly
meshed into a finite number of subregions called ele- designed implicit methods can be numerically stable over a
ments, such as tetrahedrons, quadrilaterals, and so on. wide range of integration time step values. Hence, they are
 Misra et al. 471

Figure 5. Two-dimensional ABAQUS simulation results for soft tissue deformation of the human kidney that incorporates a hyperelastic
constitutive model (Mooney-Rivlin model: C10 682.31 Pa and C01 700.02 Pa, Kim & Srinivasan, 2005) and the left side boundary nodes
are fixed, while loads are applied at the bottom and right edge nodes; (a) undeformed mesh, (b) contour plot of displacements,
(c) undeformed mesh is black, while deformed mesh is gray (red in the original version, which may be viewed as part of the
online supplemental material accompanying this article).

preferable for simulation of systems described with stiff and to run in real time. The speed of the simulation may
nonlinear differential equations. depend on the constitutive law used, the material pa-
In recent years, FE methods have been applied to rameters chosen, and the scale of the deformation. Fur-
simulate the responses of tissues and organs (for exam- ther, it is difficult to obtain the material properties of an
ple, see Figure 5). Biological tissues are anisotropic, in- inhomogeneous and anisotropic organ, which limits the
homogeneous, undergo large strains, and have nonlin- accuracy of results obtained from an FE model. Also,
ear constitutive laws, and FE techniques present an organs have anatomically complex geometries and
attractive method to numerically solve such complex boundary conditions, which are difficult to model. De-
problems. However, most commercial FE codes are op- spite these limitations and challenges, the FE technique
timized for linear elastic problems. The number of non- remains the most widely used numerical method for
linear elastic material models available in most codes is realistic modeling of surgical tool-tissue interactions.
quite limited and sensitive to small variations in material
properties. There also exist numerous challenges for
realistic simulation of organs and their application to 4 Noninvasive Soft Tissue Deformation
surgical simulators. In general, the finer the mesh in an Modeling
FE model, the more accurate the simulation. But a
greater number of elements leads to more computa- Having described the fundamental mechanics and
tional time, which hampers the ability of the simulator simulation techniques required for realistic soft tissue
472 PRESENCE: VOLUME 17, NUMBER 5

Figure 6. (a) Tools tested and measuring tool-material interaction forces during (b) and (c) deformation of rubber and (d) deformation of
bovine liver (Mahvash et al., 2002). Images printed with permission from publisher (EuroHaptics 2002).

modeling, we now begin the literature survey. We clas- homogenous materials, as stated in Equation 7. This
sify surgical tasks that do not involve tissue rupture as section describes the modeling of tool-tissue interac-
noninvasive tasks. Several modeling methods have been tions using linear elasticity-based FE methods.
considered in the literature for modeling local and In general, due to the steps involved in setting up and
global tool-tissue interactions. Some studies have also running an FE calculation, linear or nonlinear elasticity-
investigated the effect of the tool geometry on interac- based FE models cannot be simulated in real time.
tion forces. Figures 6(a), (b), (c), and (d) provide some Hence, some researchers have focused efforts on opti-
examples of the tools used to measure tool-material in- mizing FE-based computational techniques to be appli-
teraction forces (Mahvash, Hayward, & Lloyd, 2002). cable to surgical simulators. Bro-Nielsen (1998) was
It was observed that changes in tool geometry caused one of the first researchers to apply the condensation
variations in the force-deflection responses only for large method to an FE model for real-time surgical simula-
localized deformations of the material. tions. This method is based on the idea that only dis-
Most of the research presented in this section consid- placement of nodes that are in the vicinity of the tool
ered either distributed uniaxial compressive loads or need to be rendered. It was shown that nodal displace-
loads exerted by indentors, without focusing on the tool ments resulting from this method are similar to those
geometry. For the purpose of building models for surgi- obtained from conventional linear FE analysis. Compar-
cal simulators, this section categorizes the various non- ison studies were performed on an FE model of the hu-
invasive tissue modeling techniques as linear elasticity- man leg having 700 system nodes. Tensile and compres-
based and nonlinear (hyperelastic) elasticity-based FE sive loads were applied to three nodes on the calf area of
methods, and other methods that do not fall into the the leg while one edge of the leg was fixed. Cotin,
realm of continuum mechanics and/or do not use FE Delingette, and Ayache (1999) created real-time hep-
methods for simulation. tatic surgery simulations using a modified FE method
wherein the bulk of the computations were performed
during the pre-processing stage of the FE calculation.
4.1 Linear Elastic Finite Element Models
Using data from computed tomography (CT) scans,
Linear elasticity-based FE models are probably the they also built a three-dimensional anatomical model of
most widely used techniques to model tissue deforma- the liver and used linear elasticity-based modeling to
tion in surgical simulators. Motivating factors are sim- simulate its deformation. Basdogan, Ho, and Srinivasan
plicity of implementation and computational efficiency, (2001) and Kuhnapfel, Cakmak, and Maa (2000),
which enables real-time haptic rendering, since only two used linear elastic theory for developing simulators for
material constants are needed to describe isotropic and laparoscopic cholecystectomy (gallbladder removal) and
 Misra et al. 473

Figure 7. Indentation test on the Truth Cube embedded with fiducials; (a) experimental test setup, (b) CT of center vertical slice under
22% strain, (c) FE model under 22% strain (Kerdok et al., 2003). Images printed with permission from 2003 Elsevier B.V.

endoscopic surgical training, respectively. In order to done on phantom tissue with a fluid pocket. The phan-
enable real-time visual and haptic simulation, Basdogan tom tissue was deformed with a known load, while the
et al. (2001) only considered the significant vibration fluid pocket was imaged using ultrasound and the sur-
modes to compute tissue deformation, while Kuhnapfel face of the tissue was tracked using a digitizing pen.
et al. (2000) implemented the previously mentioned Kerdok et al. (2003) devised a method to measure the
condensation method. An example of a non-real-time accuracy of soft tissue models, by comparing experimen-
surgical simulation system using linear FE modeling is tal studies against FE models. They built the Truth
that of Gladilin, Zachow, Deuflhard, and Hege (2001), Cube [Figure 7(a)], which was a silicone cube embed-
who used a conventional linear FE model to simulate ded with fiducials having Youngs modulus of 15 kPa.
tissue deformations for craniofacial surgery. They found good agreement between the experimental
The models discussed above typically used assumed and simulation results for small strains (12%), where
material properties, and were not validated by compar- linear elasticity theory is valid. As expected, the linear
ing them with experimental results. Some researchers, elasticity-based FE method did not compare well against
however, have attempted to develop linear elastic FE the experimental results for large strains. Figure 7(b)
models based on experimental studies conducted on depicts the CT image for the large strain indentation
phantom or real tissue. Since noninvasive tool-tissue case. The results from the imaging studies were com-
interaction modeling is a subset of invasive surgical pro- pared to the FE model, shown in Figure 7(c). Sedef,
cedures, researchers such as DiMaio and Salcudean Samur, and Basdogan (2006) used a linear viscoelastic
(2003a), and a few others presented in Table 1 (under model where the material properties, time constants for
invasive surgical procedures) first performed indentation the relaxation function, and normalized values of shear
or other noninvasive measurements to characterize the moduli were derived from indentation experiments per-
tissue properties. Gosline, Salcudean, and Yan (2004) formed in vivo on porcine liver.
developed an FE simulation model and coupled it to the
haptic device used by DiMaio and Salcudean (2002). In
4.2 Hyperelastic Finite Element Models
Gosline et al. (2004), the authors used a linear elasticity-
based FE model to simulate organs filled with fluid. Soft tissues undergo large deformations during
Simulations were compared with experimental studies surgical procedures and the study of nonlinear solid me-
474 PRESENCE: VOLUME 17, NUMBER 5

chanics, specifically hyperelasticity, provides a frame- nomial, exponential, and logarithmic forms. Chui et al.
work for analyzing such problems. The key to studying concluded that both the Mooney-Rivlin model with
large deformation problems is the identification of an nine material constants and the combined strain energy
appropriate strain energy function. Once the strain en- of polynomial and logarithmic form with three material
ergy function is known, the constitutive stress-strain constants were able to fit the experimental data. A low-
relationships can be derived. A vast majority of the est root mean square error of 29.78 17.67 Pa was
strain energy functions used for biological soft tissues observed between analytical and experimental results for
are adapted from those used to model polymers and the tension experiments, where maximum stresses were
rubber-like materials. The Arruda-Boyce model (Arruda on the order of 3.5 kPa. Molinari, Fato, Leo, Riccardo,
& Boyce, 1993), which is based on statistical mechanics and Beltrame (2005) present a model of the scalp skin
and normally used to model rubber, has been used to to be used by plastic surgeons for pre-operative plan-
simulate palpation of breast tissues in Y. Liu, Kerdok, ning. The authors assumed a strain energy function of
and Howe (2004). X. Wu, Downes, Goktekin, and polynomial form with four parameters dependent on the
Tendick (2001) implemented the widely used Mooney- skin tissue. The maximum nodal displacement between
Rivlin model to simulate tissue deformation in the train- the simulated and experimental results was observed to
ing simulator developed by Tendick et al. (2000). They be 0.45 mm for load cases ranging from 5 N to 50 N.
introduced the concept of dynamic progressive meshing
to enable real-time computation of deformation.
4.3 Visco-Hyperelastic Finite Element
In order to develop the best possible constitutive
Models
model and add greater realism to the tissue model, sev-
eral researchers have used experimental data and elabo- Real soft tissues exhibit both viscoelastic and non-
rate setups to populate the coefficients of the strain en- linear properties. Thus, the coupling of viscoelastic and
ergy function. Carter, Frank, Davies, McLean, and hyperelastic modeling techniques results in a more real-
Cuschieri (2001) conducted several indentation tests on istic representation of soft tissues. Puso and Weiss
sheep and pig liver, pig spleen ex vivo, and human liver (1998) were the first to implement an anisotropic visco-
in vivo for the intended development of a laparoscopic hyperelastic FE model for soft tissue simulations and
surgical simulator. An exponential relationship that re- applied this technique to model the femur-medial col-
lates the stress to the stretch ratio, developed by Fung lateral ligament-tibia complex. In order to model the
(1973), was used to fit the experimental data. Davies, quasi-linear viscoelastic behavior, the authors used an
Carter, and Cuschieri (2002) conducted large and small exponential relaxation function. This was coupled with
probe indentation experiments on unperfused and per- the Mooney-Rivlin model to represent hyperelasticity of
fused pig spleen for potential use in surgical simulators. the tissue. Though simulation data were not compared
The experimental data were fitted to a hyperelastic with real tissue data and this work does not represent a
model of Neo-Hookean, Mooney-Rivlin, and exponen- surgical tool-tissue interaction model, it provides an ele-
tial forms. The goal of their study was to underscore the gant FE modeling framework for modeling soft tissues.
fact that experimental studies are required to build real- The endoscopic surgical simulator developed at ETH
istic tool-tissue interaction models, and the hyperelastic is one of the few complete systems that incorporates
model of exponential form is suitable for modeling pig continuum mechanics-based tool-tissue interaction
spleen. Hu and Desai (2004) and Chui, Kobayashi, modeling techniques, and provides realistic visualization
Chen, Hisada, and Sakuma (2004) based their model and haptic feedback in real time (Szekely et al., 2000).
on results obtained from pig liver. In Hu and Desai, the The simulator development at ETH is the culmination
authors compared results obtained from Mooney-Rivlin of many years of work and taps into the expertise of sev-
and Ogden models, while Chui et al. considered several eral engineering disciplines (Hutter, Schmitt, & Nie-
strain energy functions that were combinations of poly- derer, 2000; Szekely et al., 2000; Vuskovic, Kauer,
 Misra et al. 475

Figure 8. Formulation and results of the endoscopic simulator: (a) FE model of the human uterus containing 2000 elements. (b) Tool-tissue
interaction model used in the surgical simulator (Szekely et al., 2000). Images printed with permission from MIT Press Journals 2000 by the
Massachusetts Institute of Technology.

Szekely, & Reidy, 2000). They built a very detailed ana- model hyperelasticity. Thirteen material constants relat-
tomical model of the uterus to be simulated, followed ing to the visco-hyperelastic model were deemed suffi-
by the development of a three-dimensional homoge- cient to match the experimental data. Figure 8(a) de-
nous isotropic FE model of the organ and populated it picts the FE model and Figure 8(b) depicts tool-tissue
with real tissue material properties. Further, they de- interaction presented in Szekely et al. (2000).
signed parallel computing capability for the simulator to A few other researchers have also used visco-hyper-
function in real time and integrated a custom-built force elastic models to simulate soft tissue behavior, though
feedback device that would enable simulation of hyster- their studies are not as detailed as the work presented in
oscopy. The authors used a novel tissue aspiration Szekely et al. (2000). Work in accurate fitting of hyper-
method to capture the force-displacement relationship viscoelastic constitutive models to real tissue data is pre-
of the uterine tissue in vivo. A hyperelastic model sented below. Kim, Tay, Stylopoulos, Rattner, and
(Szekely et al., 2000) with five material constants and a Srinivasan (2003) and Kim and Srinivasan (2005) used
visco-hyperelastic model (Vuskovic et al., 2000) with data from indentation experiments on porcine esopha-
two material constants and two constants due to the gus, liver, and kidney. They fit a Blatz-form strain en-
stress relaxation function, were considered by the au- ergy function to force-displacement data obtained from
thors. Nava, Mazza, Haefner, and Bajka (2004) used quasi-static experiments, while both linear (Kelvin) and
the tissue aspiration method of Vuskovic et al. (2000) nonlinear viscoelastic models were used to fit force-time
on bovine liver and focused on modeling the precondi- data from dynamic experiments (Kim et al., 2003). A
tioning phase of soft tissue. The authors believe that Mooney-Rivlin model was used by Kim and Srinivasan.
characterizing this phenomenon can provide informa- The nonlinear viscoelastic model consisted of several
tion on the capability of the tissue to adapt to load and springs in parallel with nonlinear stiffness and was able
recover to its original configuration when unloaded dur- to match the stress relaxation curves derived from dy-
ing surgical tool and tissue interactions. A reduced poly- namic experiments. Miller (2000) and Miller and Chin-
nomial form of the strain energy function was used to zei (2002) presented a visco-hyperelastic model to sim-
476 PRESENCE: VOLUME 17, NUMBER 5

ulate the tissue response of pig brain to external loads. tion to enable virtual ultrasound display of the human
In Miller (2000), biphasic (tissue is assumed to be a thigh while providing force feedback to the user. The
mixture of a solid deformable porous matrix and a pene- model for the human thigh was composed of a mass-
trating fluid) and single phase models were evaluated, spring system whose physical parameters were based on
and the single phase model showed good correlation an earlier study conducted by dAulignac, Laugier, and
with the experimental data for up to 30% strains. The Cavusoglu (1999). The two-layer model was composed
visco-hyperelastic models considered were in terms of of a mesh of masses and linear springs, and set of non-
strain invariants and fractional powers of principal linear springs orthogonal to the surface mesh to model
stretches in Miller (2000) and Miller and Chinzei volumetric effects. The novelty of this research was that,
(2002), respectively. Both models had two independent in order to provide real-time haptic feedback to the
material parameters and one parameter relating to the user, the authors incorporated a buffer model between
stress relaxation function. In Miller (2000) theoretical the physical model and haptic device. This computa-
results were also compared with published in vivo stress- tionally simple model locally simulates the physical
strain data for Rhesus monkey liver and kidney. Real- model and can estimate contact forces at haptic update
time implementation of the simulation models has not rates. The buffer model was defined by a set of parame-
been shown. ters and was continuously adapted in order to fit the
values provided by the physical model.
In addition to continuum mechanics and FE meth-
4.4 Other Modeling Methods
ods, other innovative approaches have been developed
The primary motivation for choosing a tissue to achieve real-time performance. In order to ease the
modeling technique that is not based on linear elasticity computation burden caused by using FE-based model-
or hyperelasticity-based FE methods is to generate a ing techniques, without resorting to nonphysical meth-
computationally efficient simulation model. These spe- ods such as mass-spring-damper models, researchers
cialized models are designed for straightforward imple- have tried to implement models with two-dimensional
mentation and could be used for static and dynamic distributed elements filled with an incompressible fluid.
computation, as described in Section 3. The realism of Such models are known as the Long Element Models
tissue deformation can be compromised as a result of (LEM) and the advantage of this method is that the
such modeling simplicities, since it is difficult to relate number of the elements is one order of magnitude less
fundamental tissue properties to these models. than in an FE method based on tetrahedral or cubic
Mass-spring-damper models are the most common elements. Balaniuk and Salisbury (2002) presented the
noncontinuum mechanics-based technique used for concept of LEM to simulate deformable bodies. Their
modeling soft tissues. Organs have been modeled by approach implements a static solution for linear elastic
combining the spring-damper models, described in Sec- global deformation of objects based on Pascals princi-
tion 2.2.1, in series or parallel combination. In this case, ple and volume conservation. Using this method, it is
a set of points are linked by springs and dampers, and possible to incorporate physically based simulation of
the masses are lumped at the nodal points. Delingette, complex deformable bodes, multi-modal interactivity,
Subsol, Cotin, and Pignon (1994), Keeve, Girod, and stable haptic interface, changes in topology, and in-
Girod (1996), Koch et al. (1996), Castaneda and Coso creased graphic rendering, all done in real time. The use
(2003), and Waters (1992) are some of the studies that of static equations instead of partial differential equa-
have used mass-spring-damper models to simulate tissue tions avoids problems concerning numerical integration,
deformation, but they do not provide any information ensuring stability during simulation. Sundaraj, Men-
on the tissue properties required for the simulation. On doza, and Laugier (2002) used the concept of LEM to
the other hand, dAulignac, Balaniuk, and Laugier simulate palpation of the human thigh with a probe,
(2000) used a sophisticated apparatus for data acquisi- where the average linear elastic material constant was
 Misra et al. 477

derived from experimental studies. Sagar, Bullivant, force fields. PAFF also assumes linear elasticity for mod-
Mallinson, Hunter, and Hunter (1994) presented a de- eling soft tissues. De and Srinivasan (1999) presented
tailed and complete training system for ophthalmologi- an innovative method to model soft tissue by modeling
cal applications. Their micro-surgical training system organs as thin-walled membrane structures filled with
included a teleoperated device for the user to interact fluid. Using this technique, it was possible to model exper-
with the virtual model eye, a high-fidelity three-dimen- imental data obtained in vivo, though the authors did not
sional anatomical model of the eye, and an FE model of provide information on the simulation input parameters.
the cornea. Modeling of the collagen fibers in the cor- In order to add realism to their simulation models,
nea was done using nonlinear elastic J-shaped uniaxial some researchers have performed experimental studies
constitutive laws. Simulation tests concluded that the to populate their models with material parameters. Hu
virtual environment was able to provide graphics in real and Desai (2003) described a hybrid viscoelastic model
time. Similar to all the studies mentioned in this section, to fit the experimental results obtained during indenta-
no comparisons have been made between simulation tion experiments on pig liver. The hybrid model uses
results and actual tissue deformations during micro- linear and quadratic expressions to relate the measured
surgery in the eye. In essence, Sagar et al. (1994) used force-displacement values, which are valid for small
an FE technique for simulation, but the soft tissues were strains (up to 16% compression) and large strains (from
not modeled using linear or hyperelastic models, and 16 50% compression), respectively. The model used by
hence this work is classified in this section. the authors represents the local surface deformation of
FE modeling methods can be extremely sensitive to liver. James and Pai (2001) have achieved real-time in-
mesh resolution, so in the last decade research has been teraction by using boundary element models. If the ge-
done to avoid using meshes altogether. Such meshless, ometry, homogeneous material properties, and bound-
particle, or finite point methods share the characteristic ary conditions of the model are known, then reasonable
that there is no need to explicitly provide the connectiv- graphical update rates are achievable by precomputing
ity information between the nodes. De, Kim, Lim, and the discrete Greens functions of the boundary value
Srinivasan (2005) described a meshless technique for problem. A force interpolation scheme was used to ap-
modeling tool-tissue interactions during minimally invasive proximate forces in between time steps which allowed
surgery. They call this method the Point Collocation- for higher haptic update rate than the visual update rate.
based Method of Finite Spheres (PCMFS), wherein Inhomogeneous materials cannot be supported by the
computational particles are scattered on a domain linked boundary element analysis technique, which is a disad-
to a node. Approximation functions are defined on each vantage in applications for surgical simulation. Lang,
particle and are used to solve the differential equations Pai, and Woodham (2002) used the concept of Greens
based on linear elasticity. The PCMFS proved to be functions matrix for linear elastic deformation. The esti-
computationally superior to commercially available FE mation of the Greens function matrix was based on
packages, and performed simulations in real time. The local deformations while probing an anatomical soft-
authors are currently extending PCMFS to include non- wrist model and a plush toy. The global deformations
linear elastic properties of tissues and future work would were based on the range-flow on the objects surface. Sim-
enable users to simulate tissue cutting. The work pre- ulation and experimental results have not been compared.
sented in De et al. (2005) is based on continuum me-
chanics but does not use FE techniques for simula-
tion; hence, it is grouped in this section. In De, Lim, 5 Invasive Soft Tissue Deformation
Manivannan, and Srinivasan (2006), the authors ex- Modeling
tended the concept of PCMFS to Point-Associated Fi-
nite Field Approach (PAFF), where points are used as Almost all surgical procedures involve tissue rup-
computational primitives and are connected by elastic ture and damage, either by cutting using scissors, a
478 PRESENCE: VOLUME 17, NUMBER 5

blade, or procedures such as electro-cautery, or during area because of applications in minimally invasive percu-
operations involving needle insertion. Hence, realistic taneous procedures such as biopsies and brachytherapy.
modeling and simulation of these two operations is Research in needle insertion has examined the following
probably the most important requirement for a surgical topics: modeling and simulation of needle-tissue inter-
simulator. Further, complex but common procedures action forces, tissue deformation, deflection of the nee-
like suturing could be extrapolated from the techniques dle during insertion, path-planning of needle trajec-
developed for modeling cutting and needle insertion. tories based on tissue deformation, and devising
Modeling and simulation of invasive procedures involves experimental setups for robot-assisted needle insertion.
constantly changing boundary constraints and accurate Also, modeling of surgical cutting has focused on using
modeling of friction, which are difficult to measure. Ac- single blade scalpels or surgical scissors to model the
curate models of friction become especially important resulting soft tissue deformation. In this section we fo-
when simulating minimally invasive surgical procedures, cus on recent studies of modeling tool-tissue interaction
in which the surgeon has no direct contact with the tis- forces and tissue response during invasive procedures
sue, but manipulates the tissue via laparoscopic instru- using FE methods, while in Section 5.2 we highlight
ments. In this case, not only must sliding friction be- methods not based on continuum mechanics.
tween the instruments and the organs be accounted for,
but friction in the trocar and hinges must be modeled. 5.1.1 Linear Elastic Simulations. Tissue and
Organs are connected to bones, muscles, and/or other needle interactions have been studied by the robotics
organs via connective tissue. Hence, modeling of these community primarily for path planning of surgical pro-
connective tissues is also essential to simulate accurate cedures. DiMaio and Salcudean (2005) were the first to
response of the organ for both noninvasive and invasive develop an interactive linear elastic FE simulation model
procedures. Similar to modeling of noninvasive surgical for needle insertions in a planar environment. The simu-
procedures, linear elasticity-based FE models have been lated needle forces matched experimental data using a
the most prevalent technique for simulating invasive phantom tissue of known material properties and
operations. Very few studies have invoked nonlinear achieved real-time haptic refresh rates by using the con-
elasticity-based FE methods. Some modeling techniques densation technique (Bro-Nielsen, 1998) during pre-
that are not based on continuum mechanics are also processing. Tissue modeling techniques have also been
described in this section. implemented in steering of needles by Alterovitz, Gold-
berg, Pouliot, Taschereau, and Hsu (2003) and DiMaio
and Salcudean (2003b). Further, Goksel, Salcudean,
5.1 Finite Element Methods
DiMaio, Rohling, and Morris (2005) extended the
As mentioned earlier, the ability to model the re- work done by DiMaio and Salcudean (2003a) to inte-
sponse of soft tissue during needle insertion and/or cut- grate needle insertion simulations in three-dimensional
ting is of primary importance in the development of models. Figure 9 depicts the experimental and simula-
realistic surgical simulators. Modeling and simulation of tion work done in DiMaio and Salcudean (2003a).
invasive tissue deformation in an FE framework is signif- A two-dimensional linear FE model for needle inser-
icantly more challenging than noninvasive modeling tion during prostate brachytherapy is presented in Al-
primarily due to two factors. First, it is difficult to mea- terovitz et al. (2003). This study fine-tuned the simula-
sure the fracture toughness of inhomogeneous soft tis- tion parameters to prostate deformation results obtained
sues to accurately model the rupture process. Second, from a surgical procedure, but did not independently
invasive surgical simulation involves breaking and compare their model to data obtained by needle inser-
remeshing of nodes, which is computationally expensive tion with real or phantom tissues. Their results indicate
for reliable simulation. Nonetheless, simulation of nee- that seed placement error depends on parameters such
dle insertion through soft tissue is an active research as needle friction, sharpness, and velocity, rather than
 Misra et al. 479

Figure 9. Needle insertion and simulation modeling: (a) Probing for estimation of material properties of phantom tissue.
(b) 17 gauge epidural needle inserted into phantom tissue while motion of markers and insertion forces are recorded.
(c) FE simulation of needle insertion with small target embedded within elastic tissue (DiMaio & Salcudean, 2003a). Images
printed with permission from IEEE 2003.

patient specific parameters (tissue stiffness and com- results were presented. Chanthasopeephan, Desai, and
pressibility). Crouch, Schneider, Wainer, and Okamura Lau (2003) computed the local effective Youngs mod-
(2005) used experiments and FE modeling to show that ulus of pig liver during cutting experiments. Different
a linear elastic tissue model in conjunction with a dy- values of the effective modulus were obtained for plane
namic force function could accurately model interaction strain, plane stress, and quasi-static models, and there was a
forces and tissue deformation during needle insertion. decrease in liver resistance as the cutting speed increased.
They used a phantom tissue model with known material
properties and concluded that the accuracy of the model 5.1.2 Hyperelastic Simulations. Due to the
diminished during the relaxation phase, because soft computational burden of using FE methods for model-
tissue is viscoelastic. Hing, Brooks, and Desai (2006) ing invasive surgical procedures coupled with the diffi-
captured the different phases of interaction between culty in characterizing the nonlinear behavior of real
needle and pig liver. Using experimental data, the au- tissues during rupture, very few studies have imple-
thors estimated the linear effective modulus of the tissue mented hyperelastic models. Nienhuys and van der
sample during puncture at various speeds. Stappen (2004) used a compressible Neo-Hookean ma-
Cutting is the most common invasive surgical proce- terial model for simulating needle insertion in a three-
dure, and this operation has been modeled by some re- dimensional organ model. The study was purely based
searchers using linear elastic FE models (Picinbono, on simulations and no comparisons between real and
Lombardo, Delingette, & Ayache, 2000; W. Wu & simulation data were provided. To date, only one study
Heng, 2005). Picinbono et al. (2000) discussed the by Picinbono, Delingette, and Ayache (2003) has im-
software for a prototype laparoscopic surgical simulator plemented a nonlinear anisotropic model to simulate
which used linear extrapolation over time and position cutting of liver (hepatic resection). The anisotropic
of the interaction forces to render haptic feedback to framework is similar to the study done in Picinbono et
the user. W. Wu and Heng (2005) presented a hybrid al. (2000), and this work was extended to include hy-
condensed FE model, which consisted of operational perelasticity based on the St. Venant-Kirchhoff model.
and nonoperational regions. The authors assumed that Figure 10(a) depicts the difference in deformation be-
topological changes only occur in the operational part. tween the linear and nonlinear elasticity-based models,
The algorithm proved to be computationally efficient, while Figure 10(b) provides a screenshot simulating
but for both studies (Picinbono et al.; W. Wu & Heng), electro cautery of the liver. No validation or comparison
no comparison between simulated and experimental of the simulation model was presented.
480 PRESENCE: VOLUME 17, NUMBER 5

Figure 10. Results from work presented in Picinbono et al. (2003): (a) Comparison between linear (wireframe) versus nonlinear (solid)
elasticity-based models for same force applied to right lobe of the liver; the linear model undergoes large unrealistic deformation. (b) Simulating
hepatic resection using a nonlinear anisotropic model. Images printed with permission from 2003 Elsevier Science (USA).

5.2 Other Methods deformation was not based on actual deflection data of
the prostate, although the simulations could be per-
As discussed earlier, modeling and simulation of
formed in real time and were visually pleasing, one can-
invasive procedures requires modification of organ to-
not be certain of the realism in tissue deformation.
pology with time. Using FE methods is generally com-
Kyung, Kwon, Kwon, Kang, and Ra (2001) developed a
putationally expensive, hence, several studies have
simulator for spine needle biopsy using the voxel-based
looked at alternative modeling methods. These are pre-
haptic rendering scheme. A three-dimensional human
sented in this section. The objective of Glozman and
anatomical model was generated by segmenting images
Shoham (2004) was to formulate path planning algo-
rithms for flexible needles. Virtual springs were placed derived from CT scans or magnetic resonance imaging
orthogonal to the needle insertion axis in order to (MRI). The organs modeled in the region of the lumbar
model the needle-tissue interaction force. They did not vertebra were bones, lung, esophagus, arteries, skin,
mention the stiffness of the springs used in the simula- muscle, kidney, fat, and veins. The soft tissues were
tions, but the authors claim they can be determined ex- modeled as a series of springs, which is not realistic. The
perimentally or from pre-operative images. In order to spring stiffness was determined using needle force and
compute soft tissue deformations while simulating pros- insertion depth obtained from experimental results in a
tate brachytherapy, Wang and Fenster (2004) used a previously conducted study (Popa & Singh, 1998). The
restricted three-dimensional ChainMail method. In the skin deformation and puncture forces were modeled as a
ChainMail formulation, each volume element is linked nonlinear viscoelastic model. The simulated forces were
to its six nearest neighboring elements in the front, calculated from interactions between volume image data
back, top, bottom, left, and right. When any of the ele- and the pose of the needle.
ments is displaced beyond its defined limit (constraint In addition to developing efficient algorithms to sim-
zone), the neighboring element absorbs the movement ulate tissue rupture, the researchers presented below
due to the flexibility of the structure. The authors pro- also conducted experimental studies to populate their
posed a restricted ChainMail method by constraining models with realistic tissue data. Brett, Parker, Harrison,
the angular component of the shear constraint. The Thomas, and Carr (1997) described the design of a sur-
three-dimensional prostate image was segmented based gical needle resistance force simulator for the purpose of
on the restricted ChainMail method. Since soft tissue training and improving skills required for epidural pro-
 Misra et al. 481

cedures. The tissue model was composed of a Voigt that the small region of tool-tissue interaction was com-
mass-spring-damper model. The skin, muscular and posed of a mass-spring model, while the major part of
ligamental tissues, connective tissue and fascia, and fat the organ underwent deformation based on the linear
were modeled as nonlinear viscoelastic solid, elastic elastic FE model. Simulation results showed that this
membrane, and viscous solid, respectively. The material hybrid method was computationally efficient. However
parameters were based on fine tuning the results ob- it is very difficult to relate mass-spring parameters to
tained from porcine samples and cadavers. A similar actual material parameters.
modeling technique was used by Brett, Harrison, and All the previously mentioned models in this section
Thomas (2000), in combination with an elaborate laser- have focused on global deformation of tissue while in-
based spectroscopy technique for determining tissue teracting with a surgical tool, while the local tool-tissue
type and measuring tissue deformation. Brouwer et al. interaction is simulated as a remeshing problem ignor-
(2001) fitted an exponential relationship between the ing the energetics of cutting. In the studies presented
applied force and the stretch ratio, which were derived below, the researchers investigated and modeled local
from experimental data on various porcine abdominal tissue damage. Mahvash and Hayward (2001) at-
organs. Measurements were performed both in vivo and tempted to model cutting of soft tissues using the frac-
ex vivo during needle insertion and cutting tasks in or- ture mechanics approach. They modeled cutting of soft
der to develop a web database of tool-tissue interaction tissues as an elastic fracture undergoing plastic deforma-
models. The objective of the work done by Okamura, tion near the crack, but soft tissues in general are not
Simone, and OLeary (2004) was to model the forces linear elastic. The process of cutting was divided into
during needle insertion into soft tissue. Experimental three subtasks: deformation, cutting, and rupture,
studies were conducted ex vivo on bovine liver, with where energy exchange occurs. In the formulation,
intended applications for liver biopsy or ablation. They Mahvash and Hayward (2001) used fracture toughness
divided the forces during needle insertion into forces to describe the material property. Experimental tests
during initial puncture, forces due to friction, and forces were conducted on a potato sample and on bovine liver.
during cutting. The forces during initial puncture were In order to match experimental results with software
modeled as a nonlinear spring. The spring constants simulation results, the authors tweaked the material pa-
were obtained by curve fitting the experimental data rameters to get the best match. Further, for the experi-
and wide variation in data was observed for these con- mental studies on liver, the authors were unable to pre-
stants. A Karnopp friction model which includes both dict the different phases of fracture. Okamura, Webster,
the static and dynamic friction coefficients was used to Nolin, Johnson, and Jafry (2003) presented the Haptic
model the friction during needle insertion. Finally, the Scissors, a two-degree of freedom device that provides
cutting forces were obtained by subtracting the punc- the sensation of cutting in virtual environments by pro-
ture and friction force from the total measured force. viding force feedback. In this study, they discussed an
A clever modeling technique would incorporate real- analytical framework to model tissue cutting, and
istic tool-tissue interactions from FE models and com- showed via experimental studies that the users could not
putational efficiency from mass-spring models. Such a differentiate between the analytical model and haptic
hybrid model was presented by Cotin, Delingette, and recordings created earlier. The analytical model was a
Ayache (2000) to simulate soft tissue deformation and combination of friction, assumed material properties,
cutting. The quasi-static linear elastic FE model intro- and user motion to determine cutting forces. This sim-
duced by the authors was computationally efficient but plified model did not take into account the material
did not allow topological changes to the model. On the variations in biological tissues. The forces felt by the
other hand, the mass-spring model could simulate tear- user at the handle were assumed to be a summation of
ing and cutting in real time, but was not visually appeal- forces from friction at the scissor pivot and scissor
ing. So the authors combined the above models, such blades, and the cutting force. The data from the analyti-
482 PRESENCE: VOLUME 17, NUMBER 5

cal model did not match experimental data because the (Ottensmeyer, Kerdok, Howe, & Dawson, 2004). On
user grip force, inhomogeneous tissue properties, and the other hand, some researchers have developed novel
elastic forces in the tissue, were not modeled. Mahvash devices to measure tissue properties in vivo (Brown,
and Okamura (2005) and Mahvash, Voo, Kim, Jeung, Rosen, Sinanan, & Hannaford, 2003; Vuskovic et al.,
and Okamura (2007) applied the framework developed 2000). Brown et al. (2003) presented the modified
in Mahvash and Hayward (2001) to the Haptic Scissors. surgical graspers, while Vuskovic et al. (2000) pro-
A physically valid model would have a hyperelastic posed the tissue aspiration technique, as shown in
model describing the global tissue response while inter- Figure 11(a). Brouwer et al. (2001) described instru-
acting with the tool and the local response would be mentation to measure the soft tissue-tool forces and
governed by fracture mechanics. tissue deflection, both in vivo and ex vivo. The following
tests were performed in a pigs abdominal cavity: grasp-
ing the pig intestine wall in the longitudinal and trans-
6 Methods for Model Acquisition verse directions, indentation, needle insertion during
suturing, and cutting using scissors. Further, Ottens-
The importance of having accurate tissue models meyer (2002) described the TeMPeST 1-D (1-axis Tis-
has been recognized as a key requirement for realistic sue Material Property Sampling Tool), as shown in Fig-
and practical surgical simulators. This section presents ure 11(b), which can be used to measure linear
some of the current experimental techniques for extract- viscoelastic properties of soft tissue in vivo. TeMPeST
ing tissue properties both in vivo and ex vivo, and using 1-D is inserted laproscopically into the pig, and a wave-
invasive and noninvasive methods. Broadly, there exist form is commanded to the instrument. Data sampling
two approaches to acquire tissue properties for building takes approximately 20 seconds. Such localized mea-
surgical simulators: global and local measurement. The surement of tissue properties only provides information
choice of measurement is dictated by the intended sur- about a specific region of the organ, and for the pur-
gical simulation procedure and in turn results in the poses of modeling, local properties are usually assumed
type of experimental setup developed. The design of the to describe the behavior of the complete organ. But as
apparatus used is based on the organs structure and mentioned earlier, human organs are anisotropic and
composition, boundary conditions, and how the organ inhomogeneous, and in some cases tissue properties
is to be loaded in order to extract force-displacement vary significantly from one location to another for the
readings. same organ. Further, with localized measurement and
The most prevalent form of measuring local material modeling techniques, it is not possible to account for
properties of tissues involves indentation, uniaxial com- the organ geometry and complex boundary conditions.
pression/tensile, and/or shear experiments performed In light of the shortcomings mentioned above, some
ex vivo on a tissue sample. The applied force and tissue researchers have focused on assessing the global defor-
displacement are recorded and a constitutive law or mation of tissues to applied loads. These techniques
force-displacement relation that best fits the experimen- have typically involved placing fiducial markers on the
tal results is determined. Most of this research uses top of the tissue sample (DiMaio & Salcudean, 2003a)
phantom or ex vivo tissues, although in vivo tissues may or embedding markers (Crouch et al., 2005; Hing et
have significantly different dynamics due to variations in al., 2006; Kerdok et al., 2003) within the tissue sample.
temperature, surrounding and internal blood circula- As shown in Figures 9(a) and (b), DiMaio and Salcu-
tion, and complex boundary constraints, which are al- dean (2003a) first performed indentation, followed by
most impossible to replicate during ex vivo testing. needle insertion experiments on the phantom tissue,
Hence, some researchers have used elaborate schemes to and captured global tissue deformation using cameras.
perfuse the organ ex vivo, so as not to not compromise The displacements of the markers were tracked using
the inherent tissue properties that are observed in vivo computer vision algorithms. Dual C-arm fluoroscopes
 Misra et al. 483

Figure 11. Devices used to measure tissue properties in vivo: (a) Tissue aspiration technique (Vuskovic et al., 2000). Image printed with
permission from IEEE 2000. (b) TeMPeST 1-D with 12 mm surgical port (Ottensmeyer, 2002). Image printed with permission from Wiley-
Blackwell Publishing Ltd.

and a CT scanner were used to calculate the dislocation As an alternative to the aforementioned global mea-
of the fiducial markers in Hing et al. and Kerdok et al., surement techniques, elastography or elasticity imaging
respectively. The main limitation of this technique is is a quantitative technique to map internal tissue elastic-
that, placing markers on organs (either in vivo or ex ity. This is extremely useful in the interpretation of im-
vivo) is not practical. This is because the use of markers age data for physical modeling processes. Several elas-
in live organs might change the organ material proper- tography techniques have been developed using
ties and could possibly damage the organ. As opposed imaging modalities such as ultrasound, CT, MRI, and
to using markers, Lau, Ramey, Corso, Thakor, and optics, employing different tissue excitations, and ex-
Hager (2004) implemented an algorithm to compute tracting various parameters that provide a measure of
the surface geometry of beating pig heart in real time tissue displacement (Ophir et al., 2002). Depending on
using image intensity data. Other novel techniques that the method of tissue deformation and parameters that
do not use markers to visualize the dynamic response of are imaged, different terms are used to describe the im-
organs in vivo include attaching an ultrasound probe to ages obtained, including strain, stress, velocity, ampli-
the end of a robotically controlled laparoscopic tool tude, phase, vibration, compression, quasistatic, and
(Leven et al., 2005) and using an air pressure and strobe functional images (Gao, Parker, Lerner, & Levinson,
system to provide an image of the deformed tissue in 1996). The underlying method for estimation of tissue
real time (Kaneko, Toya, & Okajima, 2007). Leven et properties is that the organ or tissue is loaded with an
al. tested their system on liver, while Kaneko et al. de- indentor and then, using imaging, it is possible to visu-
signed their device to detect tumors in lungs. The tech- alize the internal strain in the tissues (Tonuk & Silver-
nologies presented in Lau et al., Leven et al., and Thorn, 2003; Zhang, Zheng, & Mak, 1997). One of
Kaneko et al. could be extended to measure tissue prop- the fundamental deficiencies in using elastography for
erties of organs in vivo. modeling tissues is that it is currently only possible to
484 PRESENCE: VOLUME 17, NUMBER 5

obtain Youngs modulus and Poissons ratio for the tis- most commercially available simulators, tissue material
sue, which are characteristics of linear elasticity. Turgay, parameters are tweaked based on qualitative evaluation
Salcudean, and Rohling (2006) presented two methods performed by a few surgeons, rather than actual material
to extract homogeneous and inhomogeneous tissue testing.
properties while vibrating the tissue at a spectrum of Surgical simulators with force feedback provide an
frequencies and using an ultrasound probe to capture immersive environment for training and most commer-
the tissue motion. They proposed two methods: model- cially available simulators provide haptic feedback to the
ing the tissues as a mass-spring-damper model and de- user. Reachin Technologies AB markets a simulator to
termining the transfer function from the tissue motion train for cholecystectomy performed via laparoscopic
at two separate locations. The methods were able to surgery. In order to train surgeons for endoscopic, en-
determine the tissue properties for the homogeneous dourological, percutaneous access, laparoscopic surgery,
tissue sample and only the middle layer of an inhomoge- and interventional endovascular procedures, Simbionix
neous (layered) tissue sample. USA Corp. sells GI Mentor II, URO Mentor,
As a result of the difficulties in measuring tissue prop- PERC Mentor, LAP Mentor, and ANGIO Men-
erties in vivo with previously described methods, recent tor, respectively. Immersion Medical has developed
studies have also investigated methods for online esti- numerous surgical simulation systems. The various
mation of local tissue properties during teleoperated training simulators include, CathSim AccuTouch
surgical procedures. Most of the research done in this (for vascular access and phlebotomy), Endoscopy
area is a subset of the work under environment property AccuTouch (for bronchoscopy and gastro-intestinal
estimation during telemanipulation. Duchemin, Maillet, procedures), Endovascular AccuTouch (for percutane-
Poignet, Dombre, and Pierrot (2005) used a hybrid ous transluminal angioplasty and stenting procedures),
force/position controller for robotic telesurgery in skin Hysteroscopy AccuTouch (for myomectomy and basic
harvesting procedures. With an appropriate choice of hysteroscopy skills), and Laparoscopy AccuTouch (for
gains, their controller estimated skin stiffness, friction, laparoscopic abdominal procedures). Surgical Science
and thickness. Alternatively, De Gersem, Van Brussel, Ltd. developed the LapSim system to train for abdomi-
and Vander Sloten (2005) used a Kalman filtering tech- nal keyhole surgery. ProMIS, developed by Haptica
nique to estimate the stiffness of soft materials during Inc., is used to train for minimally invasive surgery tech-
telemanipulation. Misra and Okamura (2006) proposed niques, while Mimic Technologies Inc. provides custom
an indirect adaptive estimation algorithm for estimating devices and software for the surgical simulation indus-
tissue properties while palpating the tissue. With all try. Mentice AB has the Procedicus VIST (Vascular
these online estimation methods, the challenges associ- Intervention System Trainer), Procedicus VA (Virtual
ated with local and global measurement of tissue prop- Arthroscopy), and Procedicus MIST (Minimally Inva-
erties still exist. sive System Trainer), which are used for training in pro-
cedures such as cardiac catheterization, shoulder and
knee surgery, and laparoscopic cholecystectomy, respec-
7 Commercial Surgical Simulators tively. Of these, only the Procedicus VA and Procedi-
cus MIST do not provide force feedback to the users.
In addition to the published research on modeling There exist several other commercially available surgi-
tool-tissue interactions summarized in Table 1, there cal simulators that do not provide haptic feedback, but
exist numerous commercial surgical simulators designed instead focus on training basic surgical skills (Surgical-
for the purpose of training clinicians. The methods em- SIM) or provide superior visualization capabilities
ployed by the companies for tissue modeling are not (EYESI). SurgicalSIM, developed by Medical Educa-
readily available, and there is extensive emphasis on pro- tional Technologies Inc., is used to train for general
ducing models that are visually appealing. Further, in surgical tasks on anatomically accurate models based on
 Misra et al. 485

the patients gender, size, and age. EYESI, developed proaches, meshless techniques, non-physics-based meth-
by VRmagic GmbH, is an ophthalmosurgical simulator ods, or some combination of these is the best method
that simulates stereo vision through the surgical micro- for simulating surgical tool-tissue interactions in real
scope and allows tracking of tool motion. In addition to time. The constant challenge is to develop models that
commercially available surgical simulators, research can appropriately and accurately describe the behavior
centers such as the Stanford University Medical Me- of organs, yet are computationally efficient.
dia and Information Technology develop simulators A fundamental, yet unanswered, research question is
for hospitals, and the Center for Integration of Medi- how much fidelity a surgical simulator for training needs
cine and Innovative Technology and the National in order to provide appropriate skill transfer to real pro-
Capital Area Medical Simulation Center develop sim- cedures. There is some controversy over the relevance of
ulation technologies, primarily for defense-related haptic feedback and accurate tissue modeling for skill
medical treatments. transfer, and rigorous, quantitative work needs to be
done in this area (Strom et al., 2006). In contrast, for
surgical planning the required accuracy is determined by
8 Discussion the clinical application. For example, in procedures like
biopsies of the abdominal organs, breast, and lung tis-
This review paper provides an introduction to soft sue, predicting the location of the tumor as the needle
tissue modeling, with emphasis on the relevant contin- penetrates the tissue is of primary importance, while
uum mechanics fundamentals and research to model reaching the target precisely in order to place radioac-
surgical tool and tissue interactions for the purpose of tive seeds is paramount for procedures like prostate
building surgical simulators. We broadly split surgical brachytherapy. The validation of surgical simulators is
tasks into noninvasive and invasive categories, and fur- essential to motivate their application as a method for
ther divided the modeling techniques based on linear training and pre- and intra-operative planning. Valida-
elasticity theory, nonlinear elasticity theory, and other tion techniques can be subjective (e.g., face and content
methods. We also emphasized the use of accurate organ validation) or objective (e.g., construct, concurrent, and
models by presenting some of the current methods to predictive validation) (McDougall, 2007). If we follow
acquire realistic tissue properties via experiments. Our the assumption that model fidelity and haptic feedback
emphasis was on modeling methods that employed are important, then we consider an approach in which
principles of continuum mechanics and used FE meth- we model the flow of information from the real tissue to
ods for simulation. A synopsis of the surveyed literature acquired data, the model, the rendering technique, the
is provided in Table 1, and some of the commercially haptic and/or visual display, and eventually the human
available surgical simulators are described in Section 7. user (Figure 12). We conjecture that each of these
Based on our literature search, we observed that many stages acts as a filter in which information about
researchers have done work in the area of modeling force-motion relationships is lost or transformed. For
tool-tissue interactions using linear elasticity-based FE example, the filter may be a result of the resolution of
methods. However, few studies have been published the measurement device used for gathering experimen-
using the more realistic nonlinear elasticity methods to tal data, the simulation model based on the constitutive
model invasive tool-tissue interactions. Significant con- law derived from experimental data, or simplification of
tributions have also been made using other nonstandard the model required to perform real-time haptic render-
modeling approaches. Such methods are primarily em- ing. In addition, haptic devices have their own dynamics
ployed to enable simulations to run in real time, and in and are affected by control issues such as sample-and-hold
many cases accurate physics behind tissue deformation is and quantization. Finally, the just noticeable difference
not deemed a priority. It is still an open research ques- of force perception of the human plays a vital role in
tion whether continuum mechanics-based FE ap- quantifying the necessary fidelity of the simulation.
486 PRESENCE: VOLUME 17, NUMBER 5

Figure 12. Modeling the information flow in simulator development and application. Each stage acts as a filter in which information about
force-motion relationships are lost or transformed. Images are obtained from Roberts et al. (2007), Eidgenossische Technische Hochschule,
Sacred Heart Medical Center, SensAble Technologies Inc., and Institut National de Recherche en Informatique et en Automatique. Image
corresponding to Roberts et al. (2007) printed with permission from 2007 Elsevier Ltd.

In order to accomplish surgical training and pre-oper- References

ative planning using simulators that incorporate realistic
tool-tissue interaction models along with haptic feed- Alterovitz, R., Goldberg, K., Pouliot, J., Taschereau, R., &
back, some significant challenges need to be overcome. Hsu, C. I. (2003). Needle insertion and radioactive seed
The three main hurdles that need to be conquered are: implantation in human tissues: Simulation and sensitivity
analysis. Proceedings of the IEEE International Conference on
first, the ability to obtain global responses of tool-tissue
Robotics and Automation (Vol. 2, pp. 17931799).
interactions in vivo; second, formulating a methodology
Arruda, E. M., & Boyce, M. C. (1993). Three-dimensional
to incorporate this experimental data in a continuum
constitutive model for the large stretch behavior of rubber
mechanics framework to simulate realistic deformations elastic materials. Journal of the Mechanics and Physics of Sol-
of organs; and finally, making simplifications to the ids, 41(2), 389 412.
computational model such that the simulation runs in Balaniuk, R., & Salisbury, K. (2002). Dynamic simulation of
real time but does not compromise the effective realism deformable objects using the long elements method. Pro-
of the tissue response. In spite of scientific challenges, ceedings of the 10th Symposium on Haptic Interfaces for Vir-
this literature survey and trends over the past decade tual Environments and Teleoperator Systems, 58 65.
suggest that significant strides are being made by inter- Basdogan, C., Ho, C. H., & Srinivasan, M. A. (2001). Virtual
disciplinary research teams from the robotics, mechan- environments for medical training: Graphical and haptic
simulation of laparoscopic common bile duct exploration.
ics, graphics, and medicine communities to design sur-
IEEE/ASME Transactions of Mechatronics, 6(3), 269 285.
gical simulators, and they hold great promise for
Brett, P. N., Harrison, A. J., & Thomas, T. A. (2000).
training, practice, and patient-specific pre-operative
Schemes for the identification of tissue types and boundaries
planning. at the tool point for surgical needles. IEEE Transactions on
Information Technology in Biomedicine, 4(1), 30 36.
Brett, P. N., Parker, T. J., Harrison, A. J., Thomas, T. A., &
Acknowledgments Carr, A. (1997). Simulation of resistance forces acting on
surgical needles. Proceedings of the Institution of Mechanical
The authors would like to thank Prof. Jaydev P. Desai (Uni- Engineers, Part H: Journal of Engineering in Medicine,
versity of Maryland) and Dr. Amy E. Kerdok (Harvard Uni- 211(4), 335347.
versity) for their helpful suggestions and comments. This work Bro-Nielsen, M. (1998). Finite element modeling in surgery
was supported by the Link Foundation Fellowship, National simulation. Proceedings of the IEEE, 86(3), 490 503.
Science Foundation Grant No. IIS-0347464, and National Brouwer, I., Ustin, J., Bentley, L., Sherman, A., Dhruv, N., &
Institutes of Health Grant Nos. R01-EB002004 and R01- Tendick, F. (2001). Measuring in vivo animal soft tissue
EB006435. properties for haptic modeling in surgical simulation. Medi-
 Misra et al. 487

cine Meets Virtual Reality (Vol. 98, pp. 69 74). Amster- dAulignac, D., Laugier, C., & Cavusoglu, M. C. (1999). To-
dam, The Netherlands: IOS Press. wards a realistic echographic simulator with force feedback.
Brown, J. D., Rosen, J., Sinanan, M. N., & Hannaford, B. Proceedings of the IEEE/RSJ International Conference on
(2003). In vivo and postmortem compressive properties of Intelligent Robots and Systems (Vol. 2, pp. 727732).
porcine abdominal organs. 6th International Conference on Davies, P. J., Carter, F. J., & Cuschieri, A. (2002). Mathemat-
Medical Image Computing and Computer Assisted Interven- ical modelling for keyhole surgery simulations: A biome-
tion (MICCAI) (Vol. 2879, pp. 238 245). Berlin: chanical model for spleen tissue. IMA Journal of Applied
Springer. Mathematics, 67(1), 41 67.
Carpue, J. C. (1981). An account of two successful operations Davis, A. B. (1975). Louis Thomas Jerome Auzoux and the
for restoring a lost nose, 1st ed. Birmingham, AL: The Clas- papier mache anatomical model. La Ceroplastica nella Sci-
sics of Medicine Library. enza e nellArte: atti del I Congresso Internazionale Firenze
Carter, F. J., Frank, T. G., Davies, P. J., McLean, D., & Cus- (pp. 257279). Florence, Italy.
chieri, A. (2001). Measurements and modelling of the com- De, S., Kim, J., Lim, Y.-J., & Srinivasan, M. A. (2005). The
pliance of human and porcine organs. Medical Image Anal- point collocation-based method of finite spheres (PCMFS)
ysis, 5(4), 213236. for real time surgery simulation. Computers and Structures,
Castaneda, M. A. P., & Coso, F. A. (2003). Computer simu- 83(1718), 15151525.
lation of prostate resection for surgery training. Proceedings De, S., Lim, Y.-J., Manivannan, M., & Srinivasan, M. A.
of the Annual International Conference of the IEEE Engi- (2006). Physically realistic virtual surgery using the point-
neering in Medicine and Biology (Vol. 1, pp. 11521155). associated finite field (PAFF) approach. Presence: Teleopera-
Chanthasopeephan, T., Desai, J. P., & Lau, A. C. W. (2003). tors and Virtual Environments, 15(3), 294 308.
Measuring forces in liver cutting: New equipment and ex- De, S., & Srinivasan, M. A. (1999). Thin walled models for
perimental results. Annals of Biomedical Engineering, haptic and graphical rendering of soft tissues in surgical sim-
31(11), 13721382. ulators. Medicine Meets Virtual Reality (Vol. 7, pp. 94 99).
Chui, C., Kobayashi, E., Chen, X., Hisada, T., & Sakuma, I. Amsterdam, The Netherlands: IOS Press.
(2004). Measuring forces in liver cutting: New equipment De Gersem, G., Van Brussel, H., & Vander Sloten, J. (2005).
and experimental results. Medical and Biological Engineer- Enhanced haptic sensitivity for soft tissues using teleopera-
ing and Computing, 42(6), 787798. tion with shaped impedance reflection. World Haptics Con-
Cook, R. D., Malkus, D. S., & Plesha, M. E. (1989). Concepts ference (WHC).
and applications of finite element analysis, 3rd ed. New York: Delingette, H., Subsol, G., Cotin, S., & Pignon, J. (1994). A
Wiley. craniofacial surgery simulation testbed. Proceedings of the
Cotin, C., Delingette, H., & Ayache, N. (1999). Real-time 3rd International Conference on Visualization in Biomedical
elastic deformations of soft tissues for surgery simulation. Computing (VBC) (Vol. 1, pp. 607 618).
IEEE Transaction on Visualization and Computer Graphics, Desai, C. S., & Abel, J. F. (1972). Introduction to the finite
5(1), 6273. element method: A numerical method for engineering analy-
Cotin, C., Delingette, H., & Ayache, N. (2000). A hybrid sis, 1st ed. Belmont, MA: Wadsworth Publishing Company.
elastic model for real-time cutting, deformations, and force DiMaio, S. P., & Salcudean, S. E. (2002). Simulated interac-
feedback for surgery training and simulation. The Visual tive needle insertion. Proceedings of the 10th Symposium on
Computer, 16(8), 437 452. Haptic Interfaces for Virtual Environments and Teleoperator
Crouch, J. R., Schneider, C. M., Wainer, J., & Okamura, Systems (Vol. 1, pp. 344 351). Orlando, FL.
A. M. (2005). A velocity-dependent model for needle inser- DiMaio, S. P., & Salcudean, S. E. (2003a). Needle insertion
tion in soft tissue. 8th International Conference on Medical modeling and simulation. IEEE Transactions on Robotics
Image Computing and Computer Assisted Intervention and Automation, 19(5), 864 875.
(MICCAI) (Vol. 3750, pp. 624 632). Berlin: Springer. DiMaio, S. P., & Salcudean, S. E. (2003b). Needle steering
dAulignac, D., Balaniuk, R., & Laugier, C. (2000). A haptic and model-based trajectory planning. The 6th International
interface for a virtual exam of the human thigh. Proceedings Conference on Medical Image Computing and Computer
of the IEEE International Conference on Robotics and Auto- Assisted Intervention (MICCAI) (Vol. 2879, pp. 33 40).
mation (Vol. 3, pp. 24522457). Berlin: Springer.
488 PRESENCE: VOLUME 17, NUMBER 5

DiMaio, S. P., & Salcudean, S. E. (2005). Interactive simula- H., & Szekely, G. (2006). Highly-realistic, immersive train-
tion of needle insertion models. IEEE Transactions on Bio- ing environment for hysteroscopy. In J. D. Westwood et al.
medical Engineering, 52(7), 11671179. (Eds.), Medicine meets virtual reality (Vol. 119, pp. 176
Duchemin, G., Maillet, P., Poignet, P., Dombre, E., & Pier- 181). Amsterdam: IOS Press.
rot, F. (2005). A hybrid position/force control approach Hing, J. T., Brooks, A. D., & Desai, J. P. (2006). Reality-
for identification of deformation models of skin and under- based needle insertion simulation for haptic feedback in
lying tissues. IEEE Transactions on Biomedical Engineering, prostate brachytherapy. Proceedings of the IEEE Interna-
52(2), 160 170. tional Conference on Robotics and Automation (Vol. 1, pp.
Frank, A. O., Twombly, I. A., Barth, T. J., & Smith, J. D. 619 624).
(2001). Finite element methods for real-time haptic feed- Holzapfel, G. A. (2000). Nonlinear solid mechanics: A contin-
back of soft-tissue models in virtual reality simulators. Pro- uum approach for engineering, 1st ed. New York: Wiley.
ceedings of the IEEE Virtual Reality Annual International Hu, T., & Desai, J. P. (2003). A biomechanical model of the
Symposium (VRAIS) (Vol. 1, pp. 257263). liver for reality-based haptic feedback. The 6th International
Fung, Y. C. (1973). Biorheology of soft tissues. Biorheology, Conference on Medical Image Computing and Computer
10, 139 155. Assisted Intervention (MICCAI) (Vol. 2878, pp. 75 82).
Fung, Y. C. (1993). Biomechanics: Mechanical properties of Berlin: Springer.
living tissues, 2nd ed. Berlin: Springer. Hu, T., & Desai, J. P. (2004). Characterization of soft-tissue
Fung, Y. C., & Tong, P. (2001). Classical and computational material properties: Large deformation analysis. Proceedings
solid mechanics, 1st ed. Singapore: World Scientific. of the International Symposium on Medical Simulation
Gao, L., Parker, K. J., Lerner, R. M., & Levinson, S. F. (ISMS) (Vol. 3078, pp. 28 37). Berlin: Springer.
(1996). Imaging of the elastic properties of tissueA re- Hutter, R., Schmitt, K. U., & Niederer, P. (2000). Mechani-
view. Ultrasound in Medicine and Biology, 22(8), 959 977. cal modeling of soft biological tissues for application in vir-
Gibson, S. F., & Mirtich, B. (1997). A survey of deformable tual reality based laparoscopy simulators. Technology and
modeling in computer graphics (Tech. Rep. No. TR-97-19). Health Care, 8, 1524.
Mitsubishi Electric Information Technology Center Amer- James, D., & Pai, D. K. (2001). A unified treatment of elasto-
ica. static contact simulation for real time haptics. Haptics-e,
Gladilin, E., Zachow, S., Deuflhard, P., & Hege, H. C. 2(1), 113.
(2001). A biomechanical model for soft tissue simulation in Kaneko, M., Toya, C., & Okajima, M. (2007). Active strobe
craniofacial surgery. Proceedings of the International Work- imager for visualizing dynamic behavior of tumors. Proceed-
shop on Medical Imaging and Augmented Reality (Vol. 1, ings of the International Conference IEEE on Robotics and
pp. 137141). Automation (Vol. 1, pp. 3009 3014).
Glozman, D., & Shoham, M. (2004). Flexible needle steering Keeve, E., Girod, S., & Girod, B. (1996). Craniofacial surgery
and optimal trajectory planning for percutaneous therapies. simulation. Proceedings of the 4th International Conference
The 7th International Conference on Medical Image Com- on Visualization in Biomedical Computing (VBC) (Vol. 1,
puting and Computer Assisted Intervention (MICCAI) (Vol. pp. 541546).
3217, pp. 137144). Berlin: Springer. Kerdok, A. E., Cotin, S. M., Ottensmeyer, M. P., Galea,
Goksel, O., Salcudean, S. E., DiMaio, S. P., Rohling, R., & A. M., Howe, R. D., & Dawson, S. L. (2003). Truth cube:
Morris, J. (2005). 3D needle-tissue interaction simulation Establishing physical standards for soft tissue simulation.
for prostate brachytherapy. The 8th International Conference Medical Image Analysis, 7(3), 283291.
on Medical Image Computing and Computer Assisted Inter- Kim, J., & Srinivasan, M. A. (2005). Characterization of vi-
vention (MICCAI) (Vol. 3749, pp. 827 834). Berlin: soelastic soft tissue properties from in vivo animal experi-
Springer. ments and inverse FE parameter characterization. The 8th
Gosline, A. H., Salcudean, S. E., & Yan, J. (2004). Haptic International Conference on Medical Image Computing and
simulation of linear elastic media with fluid pockets. Pro- Computer Assisted Intervention (MICCAI) (Vol. 3750, pp.
ceedings of the 12th Symposium on Haptic Interfaces for Vir- 599 606). Berlin: Springer.
tual Environments and Teleoperator Systems, 266 271. Kim, J., Tay, B. K., Stylopoulos, N., Rattner, D. W., & Srini-
Harders, M., Bajka, M., Spaelter, U., Tuchschmid, S., Bleuler, vasan, M. A. (2003). Characterization of intra-abdominal
 Misra et al. 489

tissues from in vivo animal experiments for surgical simula- Mahvash, M., & Okamura, A. M. (2005). A fracture mechan-
tion. The 6th International Conference on Medical Image ics approach to haptic synthesis of tissue cutting with scis-
Computing and Computer Assisted Intervention (MICCAI) sors. First Joint Eurohaptics Conference and Symposium on
(Vol. 2879, pp. 206 213). Berlin: Springer. Haptic Interfaces for Virtual Environment and Teleoperator
Koch, R. M., Gross, M. H., Carls, F. R., Buren, D. F. von, Systems (WHC) (Vol. 1, pp. 356 362).
Fankhauser, G., & Parish, Y. I. H. (1996). Simulating facial Mahvash, M., Voo, L., Kim, D., Jeung, K., & Okamura,
surgery using finite element models. Proceedings of the ACM A. M. (2007). Modeling the forces of cutting with scissors.
SIGGRAPH Conference on Computer Graphics and Interac- IEEE Transactions on Biomedical Engineering, 23(6),
tive Techniques (Vol. 1, pp. 421 428). New Orleans, LA. 1240 1246.
Kuhnapfel, U., Cakmak, H. K., & Maa, H. (2000). Endo- McDougall, E. M. (2007). Validation of surgical simulators.
scopic surgery training using virtual reality and deformable Journal of Endourology, 21(3), 244 247.
tissue simulation. Computers and Graphics, 24(5), 671 682. Miller, K. (2000). Biomechanics of soft tissues. Medical Sci-
Kyung, K.-K., Kwon, D.-S., Kwon, S.-M., Kang, H.-S., & Ra, ence Monitor, 6(1), 158 167.
J. B. (2001). Force feedback for a spine biopsy simulator Miller, K., & Chinzei, K. (2002). Mechanical properties of
with volume graphic model. Proceedings of the IEEE/ brain tissue in tension. Journal of Biomechanics, 35(4), 483
RSJ International on Intelligent Robots and Systems (Vol. 3, 490.
pp. 17321737). Misra, S., & Okamura, A. M. (2006). Environment parameter
Lang, J., Pai, D. K., & Woodham, R. J. (2002). Acquisition estimation during bilateral telemanipulation. Proceedings of
of elastic models for interactive simulation. International the 14th Symposium on Haptic Interfaces for Virtual Envi-
Journal of Robotics Research, 21(8), 713734. ronments and Teleoperator Systems (p. 301307). Alexandria,
Lau, W. W., Ramey, N. A., Corso, J. J., Thakor, N. V., & VA.
Hager, G. D. (2004). Stereo-based endoscopic tracking of Misra, S., Ramesh, K. T., & Okamura, A. M. (2008). Physi-
cardiac surface deformation. The 7th International Confer- cally valid surgical simulators: Linear versus nonlinear tissue
ence on Medical Image Computing and Computer Assisted models. Medicine Meets Virtual Reality (Vol. 132, pp. 293
Intervention (MICCAI) (Vol. 3217, pp. 494 501). Berlin: 295). Amsterdam, The Netherlands: IOS Press.
Springer. Molinari, E., Fato, M., Leo, G. D., Riccardo, D., & Beltrame,
Leven, J., Burschka, D., Kumar, R., Zhang, G., Blumenkranz, F. (2005). Simulation of the biomechanical behavior of the
S., Dai, X., et al. (2005). DaVinci canvas: A telerobotic sur- skin in virtual surgical applications by finite element
gical system with integrated robot-assisted, laparoscopic method. IEEE Transactions on Biomedical Engineering,
ultrasound capability. The 8th International Conference on 52(9), 1514 1521.
Medical Image Computing and Computer Assisted Interven- Nava, A., Mazza, E., Haefner, O., & Bajka, M. (2004). Ex-
tion (MICCAI) (Vol. 3749, pp. 811 818). Berlin: perimental observation and modelling of preconditioning in
Springer. soft biological tissues. The 2nd International Symposium on
Liu, A., Tendick, F., Cleary, K., & Kaufmann, C. (2003). A Medical Simulation (pp. 19).
survey of surgical simulation: Applications, technology, and Nienhuys, H.-W., & van der Stappen, F. A. (2004). A compu-
education. Presence: Teleoperators and Virtual Environ- tational technique for interactive needle insertions in 3D
ments, 12(6), 599 614. nonlinear material. Proceedings of the IEEE International
Liu, Y., Kerdok, A. E., & Howe, R. D. (2004). A nonlinear Conference on Robotics and Automation (Vol. 2, pp. 2061
finite element model of soft tissue indentation. Proceedings 2067).
of the International Symposium on Medical Simulation Ogden, R. W. (1984). Non-linear elastic deformations, 1st ed.
(ISMS) (Vol. 3078, pp. 6776). Berlin: Springer. Chichester, UK: Ellis Horwood.
Mahvash, M., & Hayward, V. (2001). Haptic rendering of Okamura, A. M., Simone, C., & OLeary, M. D. (2004).
cutting: A fracture mechanics approach. Haptics-e, 2(3), Force modeling for needle insertion into soft tissue. IEEE
112. Transactions on Biomedical Engineering, 51(10),
Mahvash, M., Hayward, V., & Lloyd, J. E. (2002). Haptic 17071716.
rendering of tool contact. Proceedings of Eurohaptics (Vol. 1, Okamura, A. M., Webster R. J., III, Nolin, J. T., Johnson,
pp. 110 115). K. W., & Jafry, H. (2003). The haptic scissors: Cutting in
490 PRESENCE: VOLUME 17, NUMBER 5

virtual environments. Proceedings of the IEEE International sium on Haptic Interfaces for Virtual Environments and Te-
Conference on Robotics and Automation (Vol. 1, pp. 828 leoperator Systems (Vol. 1, pp. 201208).
833). Spencer, A. J. M. (1972). Deformations of fibre-reinforced ma-
Ophir, J., Alam, S. K., Garra, B. S., Kallel, F., Konofagou, terials, 1st ed. Oxford, UK: Clarendon Press.
E. E., Krousko, T., et al. (2002). Elastography: Imaging the Strom, P., Hedman, L., Sarn, L., Kjellin, A., Wredmark, T.,
elastic properties of soft tissues with ultrasound. Journal of & Fellander-Tsai, L. (2006). Early exposure to haptic feed-
Medical Ultrasonics, 29(4), 155171. back enhances performance in surgical simulator training: A
Ottensmeyer, M. P. (2002). Tempest i-d: An instrument for prospective randomized crossover study in surgical resi-
measuring solid organ soft tissue properties. Experimental dents. Surgical Endoscopy, 20(9), 13831388.
Techniques, 26(3), 48 50. Sundaraj, K., Mendoza, C., & Laugier, C. (2002). A fast
Ottensmeyer, M. P., Kerdok, A. E., Howe, R. D., & Dawson, method to simulate virtual deformable objects with force
S. L. (2004). The effects of testing environment on the vis- feedback. The 7th International Conference on Automation,
coelastic properties of soft tissues. Proceedings of Medical Robotics, Control, and Vision (ICARCV) (Vol. 1, pp. 413
Simulation: International Symposium (ISMS) (Vol. 3078, 418).
pp. 9 18). Berlin: Springer. Szekely, G., Brechbuhler, C., Dual, J., Enzler, R., Hug, J.,
Picinbono, G., Delingette, H., & Ayache, N. (2003). Non- Hutter, R., et al. (2000). Virtual reality-based simulation of
linear anisotropic elasticity for real-time surgery simulation. endoscopic surgery. Presence: Teleoperators and Virtual En-
Graphical Models, 65(5), 305321. vironments, 9(3), 310 333.
Picinbono, G., Lombardo, J.-C., Delingette, H., & Ayache, Tendick, F., Downes, M., Goktekin, T., Cavusoglu, M. C.,
N. (2000). Anisotropic elasticity and force extrapolation to Feygin, D., & Wu, X. (2000). A virtual environment test-
improve realism of surgery simulation. Proceedings of the bed for training laparoscopic surgical skills. Presence: Teleop-
IEEE International Conference on Robotics and Automation erators and Virtual Environments, 9(3), 236 255.
(Vol. 1, pp. 596 602). Tonuk, E., & Silver-Thorn, B. (2003). Nonlinear elastic mate-
Popa, D. O., & Singh, S. K. (1998). Creating realistic force rial property estimation of lower extremity residual limb
sensations in a virtual environment: Experimental system, tissues. IEEE Transactions on Neural Systems and Rehabili-
fundamental issues and results. Proceedings of the IEEE In- tation Engineering, 11(1), 4353.
ternational Conference on Robotics and Automation (Vol. 1, Turgay, E., Salcudean, S., & Rohling, R. (2006). Identifying
pp. 59 64). the mechanical properties of tissue by ultrasound strain im-
Puso, M. A., & Weiss, J. A. (1998). Finite element implemen- aging. Ultrasound in Medicine and Biology, 32(2), 221235.
tation of anisotropic quasi-linear viscoelasticity using a dis- Vuskovic, V., Kauer, M., Szekely, G., & Reidy, M. (2000).
crete spectrum approximation. ASME Journal of Biome- Realistic force feedback for virtual reality based diagnostic
chanical Engineering, 120(1), 6270. surgery simulators. Proceedings of the IEEE International
Roberts, J. C., Merkle, A. C., Biermann, P. J., Ward, E. E., Conference on Robotics and Automation (Vol. 2, pp. 1592
Carkhuff, B. G., Cain, R. P., et al. (2007). Computational 1598).
and experimental models of the human torso for non-pene- Wang, X., & Fenster, A. (2004). A virtual reality based 3D
trating ballistic impact. Journal of Biomechanics, 40(1), 125 real-time interactive brachytherapy simulation of needle in-
136. sertion and seed implantation. IEEE International Sympo-
Sagar, M. A., Bullivant, D., Mallinson, G. D., Hunter, P. J., sium on Biomedical Imaging: From Nano to Macro (Vol. 1,
& Hunter, I. W. (1994). A virtual environment and model pp. 280 283).
of the eye for surgical simulation. Proceedings of the 21st An- Waters, K. (1992). Physical model of facial tissue and muscle
nual Conference on Computer Graphics and Interactive Tech- articulation derived from computer tomography data. Pro-
niques (SIGGRAPH) (Vol. 1, pp. 205212). ceedings of SPIEThe International Society for Optical Engi-
Satava, R. M. (2001). Accomplishments and challenges of sur- neering (Vol. 1808, pp. 574 583).
gical simulation. Surgical Endoscopy, 15(3), 232241. Wu, W., & Heng, P. A. (2005). An improved scheme of an
Sedef, M., Samur, E., & Basdogan, C. (2006). Visual and interactive finite element model for 3D soft-tissue cutting
haptic simulation of linear viscoelastic tissue behaviour and deformation. The Visual Computer, 21(8 10), 707
based on experimental data. Proceedings of the 14th Sympo- 716.
 Misra et al. 491

Wu, X., Downes, M. S., Goktekin, T., & Tendick, F. (2001). friction and large deformation. Medical Engineering & Phys-
Adaptive nonlinear finite elements for deformable body sim- ics, 19(6), 512517.
ulation using dynamic progressive meshes. Computer Zhukov, L., & Barr, A. H. (2003). Heart-muscle fiber re-
Graphics Forum, 20(3), 349 358. construction from diffusion tensor MRI. Proceedings of
Yamada, H. (1970). Strength of biological materials, 1st ed. the IEEE Visualization Conference (Vol. 1, pp.
Baltimore: Williams & Wilkins. 597 602).
Zhang, M., Zheng, Y. P., & Mak, A. F. (1997). Estimating Zienkiewicz, O. C., Taylor, R. L., & Zhu, J. Z. (2005). The
the effective Youngs modulus of soft tissues from indenta- finite element method: Its basis and fundamentals, 6th ed.
tion testsNonlinear finite element analysis of effects of Oxford, UK: Elsevier Butterworth-Heinemann.